REVIEW ARTICLE Genetic approaches to the cellular functions of polyamines in mammals Juhani Ja¨ nne, Leena Alhonen, Marko Pietila¨ and Tuomo A. Keina¨nen A.I. Virtanen Institute for Molecular Sciences, University of Kuopio, Kuopio, Finland The polyamines putrescine, spermidine and spermine are organic cations shown to participate in a bewildering num- ber of cellular reactions, yet their exact functions in inter- mediary metabolism and specific interactions with cellular components remain largely elusive. Pharmacological inter- ventions have demonstrated convincingly that a steady supply of these compounds is a prerequisite for cell prolif- eration to occur. The last decade has witnessed the appear- ance of a substantial number of studies, in which genetic engineering of polyamine metabolism in transgenic rodents has been employed to unravel their cellular functions. Transgenic activation of polyamine biosynthesis through an overexpression of their biosynthetic enzymes has assigned specific roles for these compounds in spermatogenesis, skin physiology, promotion of tumorigenesis and organ hyper- trophy as well as neuronal protection. Transgenic activa- tion of polyamine catabolism not only profoundly disturbs polyamine homeostasis in most tissues, but also creates a complex phenotype affecting skin, female fertility, fat depots, pancreatic integrity and regenerative growth. Transgenic expression of ornithine decarboxylase antizyme has sugges- ted that this unique protein may act as a general tumor suppressor. Homozygous deficiency of the key biosynthetic enzymes of the polyamines, ornithine and S-adenosyl- methionine decarboxylase, as achieved through targeted disruption of their genes, is not compatible with murine embryogenesis. Finally, the first reports of human diseases apparently caused by mutations or rearrangements of the genes involved in polyamine metabolism have appeared. Keywords: antizyme; ornithine decarboxylase; putrescine; spermidine/spermine N 1 -acetyltransferase; spermidine; sper- mine; transgenic mouse; transgenic rat. Introduction The cellular functions of the natural polyamines (putrescine, spermidine and spermine) are still largely unknown, although a vast number of studies have shown that these polycationic compounds are crucial to the growth and proliferation of mammalian cells. Pharmacological approa- ches are applied typically in studies aimed to unravel their functions in cellular metabolism and, admittedly, much valuable information has been generated with the use of specific inhibitors of polyamine biosynthesis. However, the last decade has produced a substantial number of experi- mental studies in which genetic engineering of polyamine metabolism has been used as a tool to elucidate their cellular functions. Studies with genetically engineered mice and rats have not only brought entirely new information about the involvement of polyamines in various physiological and pathophysiological processes but they have likewise chal- lenged some of the conventional wisdoms. Mainly, four different approaches have been applied in the genetic engineering of experimental animals: (a) activation of polyamine biosynthesis through the overexpression of their biosynthetic enzymes; (b) activation of polyamine catabol- ism through the overexpression of the enzymes involved in their catabolism; (c) transgenic expression of ornithine decarboxylase (ODC) antizyme, a protein inhibiting ODC activity and facilitating its degradation and (d) gene- disruption technology applied both to the biosynthetic and catabolic enzymes. Polyamine metabolism Figure 1 outlines the metabolism of the polyamines in a mammalian cell. Two amino acids, L -ornithine and L -methionine, are the primary precursors of the poly- amines. L -Ornithine is cleaved from L -arginine by mito- chondrial arginase II [1] or derived from the diet and L -methionine is first converted to S-adenosyl- L -methionine (AdoMet). Both ornithine and AdoMet are subsequently decarboxylated by two cytosolic decarboxylases, namely ornithine decarboxylase (ODC) and AdoMet decarboxy- lase (AdoMetDC). The former reaction yields putrescine Correspondence to J. Ja ¨ nne, A.I. Virtanen Institute for Molecular Sciences, University of Kuopio, PO Box 1627, FIN-70211, Kuopio, Finland. Fax: + 358 17 163025, Tel.: + 358 17 163049, E-mail: Juhani.Janne@uku.fi Abbreviations: ODC, ornithine decarboxylase; AdoMetDC, S-adeno- sylmethionine decarboxylase; dcAdoMet, decarboxylated adeno- sylmethionine; SSAT, spermidine/spermine N 1 -acetyltransferase; PAO, polyamine oxidase; SMO, spermine oxidase; DENSPM, N 1 ,N 11 -diethylnorspermine; DFMO, 2-difluoromethylornithine; MGBG, methylglyoxal bis(guanylhydrazone); NMDA, N-methyl- D -aspartate; GABA, c-aminobutyric acid; eIF5A, eukaryotic initiation factor 5 A. (Received 12 December 2003, revised 19 January 2004, accepted 22 January 2004) Eur. J. Biochem. 271, 877–894 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04009.x (NH 2 CH 2 CH 2 CH 2 CH 2 NH 2 ) and the latter reaction decarboxylated AdoMet (dcAdoMet). DcAdoMet do- nates its aminopropyl group either to putrescine in a reaction catalyzed by a transferase, spermidine synthase to yield spermidine (NH 2 CH 2 CH 2 CH 2 NHCH 2 CH 2 CH 2 CH 2 NH 2 ), or to spermidine catalyzed by a separate transferase, spermine synthase to yield spermine (NH 2 CH 2 CH 2 CH 2 NHCH 2 CH 2 CH 2 CH 2 NHCH 2 CH 2 CH 2 NH 2 ). As the decarboxylation and propylamine transferase reactions are practically irreversible, an entirely distinct system exists to convert the higher polyamines back to putrescine. This system utilizes two different enzymes, a cytosolic spermidine/spermine N 1 -acetyltransferase (SSAT) [2] and a peroxisomal flav- oprotein polyamine oxidase (PAO) [3]. As PAO strongly prefers acetylated polyamines as the substrates [3,4], SSAT is the rate-controlling enzyme in this backconver- sion pathway [4]. As indicated in Fig. 1, spermine can be either monoacetylated or diacetylated [5] by SSAT. As seen, diacetylation of spermine would not require a re- entry of spermidine back to the peroxisome but putres- cine would be formed via N 1 -acetylspermidine from diacetylated spermine (Fig. 1). In addition to putrescine and spermidine, the PAO reaction also yields hydrogen peroxide and acetaminopropanal. Working with SSAT- deficient mouse embryonic stem cells, we found that SSAT is absolutely necessary for the conversion of spermidine to putrescine while spermine is readily converted to spermidine in the absence of SSAT [6]. The conversion of spermine back to spermidine in the absence of SSAT activity is obviously attributable to a recently discovered amine oxidase, which, when first cloned was thought to be PAO [7], but was soon identified as a novel spermine oxidase (SMO) [8]. Unlike PAO, SMO strongly prefers spermine as the substrate over its acetylated derivatives and spermidine is not a substrate at all [8,9]. Figure 1 likewise highlights an important Ôside trackÕ of spermidine metabolism, namely it serves as a precursor for hypusine synthesis. This unusual amino acid, derived from the aminobutyl moiety of spermidine, represents an integral part of eukaryotic initiation factor 5 A (eIF5A) [10] that is essential for eukaryotic cell proliferation [11]. A further unique protein, ODC antizyme, is intimately involved in the metabolism of the polyamines. Antizyme was initially discovered in the late 1970s as a protein inhibitor of ODC [12–14]. Subsequently, it became obvious that the antizyme not only inhibited ODC activity but it also facilitated its degradation by targeting ODC to 26S proteasome [15,16]. It also appears that the antizyme is responsible for the feedback inhibition of polyamine transport [17]. The regulation of ODC antizyme expres- sion is unique as polyamines directly induce ribosomal frameshifting in decoding antizyme resulting in the formation of full-length functional protein [18,19]. Recently, a nuclear localization has been described for antizyme (and ODC) during mouse development [20] and antizyme appears to have nuclear export signals [21]. Recent studies have likewise indicated that antizyme interacts in the nucleus with the transcription factor Fig. 1. The metabolism of the polyamines. ODC, ornithine decarboxylase; Spd, spermidine; AdoMetDC, S-adenosylmethionine decarboxylase; dcAdoMet, decarboxylated AdoMet; Spm, spermine; SSAT, spermidine/spermine N 1 -acetyltransferase; PAO, polyamine oxidase; SMO, spermine oxidase; eIF5A, eukaryotic initiation factor 5 A. The superscripts indicate the genetic modification of the genes: TG, transgenic; KO, knockout; DN, dominant negative mutation. 878 J. Ja ¨ nne et al. (Eur. J. Biochem. 271) Ó FEBS 2004 Smad1 and with HsN3, a proteasome subunit [22]. The complex then appears to recruit SNIP 1, the repressor of CBP/p300 [23]. The latter apparently implies that the antizyme has functions beyond the regulation of ODC activity/turnover and polyamine transport. There are several recent review articles dealing in more detail with polyamine metabolism, their putative functions and polyamine-related pharmacological/clinical approaches [24–27]. Putative cellular functions of the polyamines Polyamines are organic cations that are positively charged under physiological conditions. Thus, they are expected to interact with negatively charged molecules, such as nucleic acids, phospholipids, etc., within the living cells. What makes them different from divalent cations, for instance? There are two fundamental differences between the natural polyamines and divalent cations, such as Mg 2+ and Ca 2+ . (a) The positive charges in the polyamines are differentially spaced within a flexible carbon backbone and hence electrostatic interactions with other cellular components, most notably with negatively charged macromolecules, in all likelihood are more flexible than those exerted by divalent cations. (b) As described above, the polyamines possess extremely sophisticated metabolic machinery for the regulation and maintenance of their intracellular homeo- stasis. Of course, one can argue that the latter likewise holds true for divalent cations in terms of differential cellular compartmentalization. There are more than 1500 research reports describing effects exerted by the polyamines in great diverse experimental systems in vitro. If not all, most of these experiments may not be relevant as regards the true cellular functions of the polyamines, as no one knows the concentrations of free polyamines in a living cell. Especially the higher polyamines, spermidine and spermine while present in millimolar concentrations in the cell are expected to be tightly bound to negatively charged cellular compo- nents and structures leaving only an extremely tiny fraction of them reactive. Thus, to conclude anything about the functions of the polyamines based on biological effects exerted by them in vitro is a hopeless task without any solid ground. As regards the physiological functions of the polyamines, however, experimental evidence exists that assigns specific roles to the polyamines in general and to individual polyamines in particular. The polyamines, especially sper- midine and spermine, interact with DNA with reasonable specificity. In fact, they can alter the structure of DNA, such as B to Z conversion, and are thus likely to affect the function of DNA [24]. An example of an extremely specific interaction between the higher polyamines and polynucleo- tides is the ribosomal frameshifting induced by the poly- amines in decoding of ODC antizyme [18]. Polyamines appear to have an indispensable role in cell proliferation, as specific inhibition of their biosynthesis invariably halts the growth of mammalian cells. This likewise applies to polyamine depletion achieved by an activation of their catabolism. The biosynthetic and catabolic enzymes of the polyamines have become a meaningful target for cancer chemotherapy. However, the impressive results obtained in cell cultures and even with tumor-bearing animal models have not fully translated to clinical practice apparently due to the sophisticated compensatory mechanisms aimed at maintaining polyamine homeostasis [25,26]. In fact, poly- amines may have a dual role in cellular functions by promoting cell growth or inducing apoptosis when they occur in excess [24]. An inhibition of polyamine biosynthesis has met much greater success in the treatment of certain parasitic diseases, such as African sleeping sickness, under clinical conditions [24]. A further function of the poly- amines, specific to spermidine, is the formation of hypusine, an integral component of the eIF5A [10]. As the latter factor is essential for the proliferation of mammalian cells [11], it is often difficult to judge whether spermidine depletion- induced growth arrest is attributable to the polyamine itself or whether it is related to the shortage of hypusine to form a functional translation initiation factor. Finally, the poly- amines specifically interact with certain ion channels, such as N-methyl- D -aspartate receptor, inward rectifying potassium channels and voltage-dependent Ca 2+ channels [24]. As these interactions occur at such low (nanomolar) concen- trations, it is highly likely that they are relevant also under conditions in vivo. Genetic engineering has been directed to almost every single reaction of polyamine metabolism in transgenic animals and embryonic stem cells. Individual enzymes include arginase II, ODC, antizyme, AdoMetDC, spermi- dine synthase, spermine synthase and SSAT the genes of which have been either generally or tissue-specifically overexpressed or disrupted. The following sections will describe in detail the consequences of activated polyamine biosynthesis, activated polyamine catabolism, antizyme overexpression and disruption of individual enzyme genes. A short recent review also covers some of the transgenic mouse models [28]. Activation of polyamine biosynthesis in transgenic rodents Overexpression of ornithine decarboxylase Polyamine homeostasis. The first animal with genetically engineered polyamine metabolism was a transgenic mouse overexpressing the human ODC gene under its own promoter [29]. The transgene was overexpressed practically in all tissues of the transgenic mice in a position-independent and gene copy number-dependent fashion [30]. While the parenchymal tissues displayed moderately elevated ODC activity, some tissues, such as testis and brain, showed an enzyme activity that was 20–80 times higher than that in the respective tissues of their nontransgenic littermates. The high activity of ODC greatly expanded tissue pools of putrescine, especially in testis and brain [31], but, with the possible exception of testis, was not reflected in any alterations of tissue levels of spermidine and spermine [31]. The fact that testis and brain showed the most expanded putrescine pools in the transgenic animals may be attribut- able to their less permeable blood/tissue barriers than in other tissues. This apparent block between putrescine and the higher polyamines is especially puzzling as analyses of SSAT and PAO activities and urinary content of the polyamines gave no signs indicative of an activation of polyamine catabolism [31] or their enhanced urinary Ó FEBS 2004 Genetic approaches to polyamine functions (Eur. J. Biochem. 271) 879 excretion [32] in the transgenic animals. An elevation of tissue contents of L -ornithine through the inhibition of ornithine transaminase only further expanded putrescine pools but not those of spermidine or spermine [32], indicating that the tissue ornithine pool became rate-limiting under these conditions. These results led us to conclude that one of the major functions of the polyamine homeostatic system in nondividing (or nonstressed) mammalian cells is to prevent an excessive accumulation of the higher poly- amines spermidine and spermine, in fact, a view already expressed by Davis et al. [33] in their review article. If correct, this would require some sort of physical or chemical sequestration of putrescine under these conditions. Spermatogenesis. While establishing the first transgenic ODC overexpressing transgenic mouse line (UKU2), we generated a transgenic founder male displaying extremely high testicular ODC activity. This male turned out to be infertile and subsequent histological examination of testes revealed greatly reduced amount of germinal epithelium and total absence of ongoing spermatogenesis resembling a syndrome causing infertility in man and known as ÔSertoli cells-only syndromeÕ [29]. A closer examination of the members of the UKU2 line revealed a significant decrease also in their sperm count [29]. Our subsequent studies indicated that an enhanced testicular ODC activity has a dual effect on spermatogenesis; a moderately enhanced putrescine accumulation stimulates mitotic DNA synthesis while excessive accumulation of the diamine inhibits meiotic DNA synthesis, particularly upon advancing of age [34], ultimately leading to male infertility [29]. Highly relevant to these findings is a recent discovery of a testis germ cell- specific ODC antizyme (antizyme 3) [35]. The expression of this antizyme is strictly restricted to testis and it is expressed stage-specifically in postmeiotic germ cells, i.e. during late spermatogenesis [35]. As indicated by our results with ODC transgenics, excessive putrescine accumulation during late spermatogenesis appears to be detrimental to the germ cells and, hence, antizyme 3 apparently functions to limit any ODC expression to the early spermatogenesis. The trans- gene-derived ODC seemingly ÔoverwhelmsÕ the normally occurring antizyme and leads to an excessive accumulation of putrescine in late-stage germ cells [35,36]. Spontaneous tumorigenesis. Elevated ODC activity and expanded pools of the polyamines are commonly associated with tumorigenesis and a role of an oncogene-like protein has been assigned to ODC [26,37]. Under cell culture conditions, forced expression of ODC appears to result in malignant transformation [37,38] and cells overproducing ODC are able to form tumors in nude mice [39]. We subjected the ODC-overexpressing transgenic mice to a long-term survival study in order to assess whether very high constitutive tissue ODC activity would predispose the animals to an enhanced general tumorigenesis. At two years of age, these animals still displayed 20–50 times higher tissue ODC activity than their syngenic littermates, yet macro- scopic and microscopic examination of organs did not reveal any difference between syngenic and transgenic mice as regards spontaneous tumor incidence [40]. These findings are supported indirectly by results obtained with growth hormone overexpressing transgenic mice, which, in addition to constitutively elevated circulating growth hormone levels, showed enhanced ODC activity in liver and some other tissues but did not display any signs of malignant transfor- mation even at advanced age [41]. A more recent study with transgenic mice carrying mammary tumor virus long- terminal repeat-driven ODC cDNA suggests an increased incidence of spontaneous tumors in the transgenic animals [42]. The tumors included mammary carcinomas, intestinal adenocarcinoma and a vascular liver tumor, however, a direct connection to ODC is questionable as the listed transgenic tissues displayed lower ODC activity than the corresponding tissues of syngenic mice. Another strange feature is the fact that among 17 nontransgenic animals, no pathological findings were observed at 2 years of age, which is close to the end of the life-span of a mouse [42]. According to a recent report, spontaneous tumor incidence in old mice of CH3 background (used in the cited study) is 40% [43]. Skin. Transgenic mice overexpressing human ODC under its own promoter did not show any macroscopic or microscopic skin abnormalities, yet these animals appeared to be more sensitive than their syngenic littermates to developing skin papillomas in response to the two-stage chemical (initiation and promotion) skin tumorigenesis [44]. Targeted (using bovine keratin promoter) overexpression of truncated ODC in the skin of transgenic mice caused a number of phenotypic abnormalities including early and permanent loss of hair, excessive skin wrinkling, development of dermal follicular cysts, enhanced nail growth and spontaneous tumor devel- opment [45]. The fact that hair loss was causally related to ODC overexpression and putrescine accumulation was convincingly proved by experiments showing that early administration of 2-difluoromethylornithine (DFMO), an irreversible inhibitor of ODC [46], prevented the hair loss and partially normalized skin histology [47]. The same authors likewise showed that hair follicle-targeted overexpression of ODC not only predisposed the transgenic animals to skin tumorigenesis but the tumors developed in response to a single carcinogen application without a subsequent tumor promotion [48]. The use of DFMO reversibly blocked the formation of squamous papillomas in the transgenic animals, which led the authors to conclude that polyamines, most notably putrescine, control the development and mainten- ance of neoplastic phenotype [49]. The overexpression of ODC appears to co-operate with v-Ha-ras oncogene as doubly transgenic mice carrying both keratin promoter- driven ODC transgene and v-Ha-ras developed spontaneous skin tumors unlike the singly transgenic animals [50]. In addition to chemical carcinogenesis, transgenic mice over- expressing ODC under the control of keratin promoter were also sensitized to photocarcinogenesis as indicated by the more rapid development of skin tumors in these animals in comparison with their syngenic littermates [51]. As in the case of chemical carcinogenesis, the development of tumors in response to the ultraviolet radiation was completely preven- ted by DFMO [51]. The role of ODC in skin tumorigenesis has also been approached by generating transgenic mouse line expressing keratin promoter-driven truncated dominant-negative ODC mutant [52]. In spite of an inhibition of wild-type ODC expression in short-term experiments, these animals formed as many tumors as controls in response to the two-stage skin 880 J. Ja ¨ nne et al. (Eur. J. Biochem. 271) Ó FEBS 2004 tumorigenesis protocol. A plausible explanation for the failure to inhibit tumorigenesis is the competition of the mutant ODC for the binding of the antizyme and releasing thewild-typeenzyme[52]. The mechanisms whereby ODC overexpression enhances skin tumorigenesis are largely unknown. Two recent reports link ODC overexpression to histone acetylation as both histone acetyltransferase and deacetylase activities were elevated in the skin of the ODC transgenic animals and histones were hyperacetylated in cultured skin cells over- expressing ODC [53,54]. Skin tumors obtained from doubly transgenic ODC/Ras mice displayed an exceptionally high histone acetyltransferase activity [54]. These changes were fully reversible by DFMO treatment. These findings may imply that elevated intracellular polyamines can influence the chromatin organization and possibly alter specific gene expression to promote tumor progression [54]. As further regards skin, it may be worth of mentioning that an overexpression of arginase I in the enterocytes of transgenic mice elicits arginine deficiency that affects skin, muscle and lymphoid development, however, in the absence of altered polyamine tissue pools [55]. Central nervous system. The role of the polyamines in normal and pathological brain is not only under active research but the existing views of their roles are highly conflicting. A vast number of experiments have shown that brain insults, either physical or chemical, inevitably activate the biosynthesis of the polyamines through an induction of ODC and concomitant accumulation of putrescine. This phenomenon is mostly understood in terms that the induction of ODC and the enhanced accumulation of putrescine is causally related to the neuronal damage rather than representing an adaptive response [56,57]. Spermidine and spermine interact as agonists with the N-methyl- D -aspartate (NMDA) receptor [58,59], a prolonged activa- tion of which could be responsible for neuronal damage [56]. Unlike the higher polyamines, putrescine is believed to act as a weak antagonist of the NMDA receptor [59]. The experiments with transgenic mice and rats overexpressing ODC have generated new information strongly suggesting that an enhanced accumulation of putrescine in brain is a neuroprotective measure rather than a cause of neuronal damage. As indicated earlier, transgenic mice overexpressing ODC show the greatest expansion of the putrescine pool in brain and testis. In the long-term survival experiment, we examined the transgenic animals and their syngenic litter- mates at 2 years of age and found no macroscopic or microscopic signs of neuronal degeneration in the transgenic brains [40]. This means that life-long constitutive over- expression of ODC and enhanced accumulation of putres- cine in the brain is tolerated with no harmful consequences. Consequent experiments with the ODC overexpressing mice indicated that these animals showed a significantly elevated seizure threshold to both chemical (pentylene- tetrazol) and physical (electroshock) stimuli and impaired performance in spatial learning and memory tests. The elevated seizure threshold was not due to any changes in the brain levels of the two major neurotransmitter amino acids, glutamate and c-aminobutyric acid [60]. Mg 2+ is a well known voltage-dependent, physiological blocker of the glutamate-mediated excitatory currents inhibiting ionic conductance through the NMDA channel [61]. Thus, elevated free Mg 2+ could potentially block the NMDA receptor, yet our studies revealed that free Mg 2+ was significantly lowered (40%) in the brain of transgenic animals [62]. Taken together, these results suggest that endogenous putrescine may play a physiologically relevant role at the NMDA receptor as these receptors have a well documented role in the induction of seizure activity [63] and mediating spatial encoding [64]. The finding indicating that the transcript levels of several neurotrophins were elevated in the brain of the transgenic animals may likewise contribute to the apparent neuroprotection [65]. The view that elevated brain putrescine offers neuropro- tection, or at least is not neurotoxic, was supported by a series of studies where ODC overexpressing transgenic mice and rats were subjected to cerebral ischemia. Transgenic mice suffering from incomplete forebrain ischemia due to the occlusion of common carotid arteries did not show any signs indicative of putrescine neurotoxicity as judged by changes of energy metabolism (assessed with the aid of nuclear magnetic resonance spectroscopy), induction of immediate early genes and the extent of hippocampal necroses [66]. Similar results were obtained with a transgenic rat model for ODC overexpression after permanent middle cerebral artery occlusion [67]. A comparison of ODC overexpressing transgenic rats with syngenic and DFMO- treated rats after transient middle cerebral occlusion indi- cated that the ischemia-reperfusion damage developed more slowly and the infarct volumes were smaller in the trans- genic animals [68,69]. These studies clearly indicate that an induction of ODC and the concomitant accumulation of brain putrescine are adaptive responses to noxious insults and do not enhance the lesion development. Cardiac hypertrophy. Agents that cause cardiac hyper- trophy are known to activate polyamine biosynthesis and elevate their cardiac levels [70,71]. Although cardiac hyper- trophy in response to b-adrenergic agonists can be prevented by a specific inhibition of ODC by DFMO [71,72], it is by no means clear whether an enhanced polyamine biosynthesis and accumulation per se can cause cardiac hypertrophy. This issue was addressed by generating transgenic mice with targeted overexpression of ODC in the heart. Using truncated ODC driven by a-myosin-heavy-chain promoter, a more than 1000-fold overexpression was achieved [73]. Cardiac putrescine pool was expanded by a factor of 50 and that of spermidine by a factor of four while spermine content was only slightly elevated [73]. The enormous ODC activity apparently depleted tissue ornithine, as substantial amounts of lysine-derived cadaverine accumulated in the heart. Even though the altered polyamine pools in transgenic animals did not lead to hypertrophic phenotype, this condition co-operated with b-adrenergic stimulation resulting in severe, sometimes fatal, cardiac hypertrophy compared with only mild hypertrophy in the similarly treated nontransgenic littermates [73]. It is noteworthy that in this study, ODC overexpression resulted in substantial expansion of tissue spermidine pool too that is not often observed in other ODC overexpressing transgenic models. Polyamines are known to be modulators of inward rectifying K channels (K ir channels) spermine being 100-fold Ó FEBS 2004 Genetic approaches to polyamine functions (Eur. J. Biochem. 271) 881 and spermidine 10-fold more potent blockers of the channels than putrescine or Mg 2+ [74–76]. Studies on inward rectification properties of cardiomyocytes isolated from transgenic mice with heart-targeted ODC overexpres- sion, unexpectedly revealed that, in spite of massive over- accumulation of putrescine and cadaverine, the rectification properties were essentially unaltered [76]. Two explanations for this finding were offered: (a) these diamines did not significantly contribute to the rectification or (b) their free concentrations were not altered despite the massive rise in total levels. Interestingly, the authors reached the conclusion proposing that most of the putrescine (and cadaverine) is not free but is sequestered within the cell [76]. If correct, this would also explain the commonly observed biosynthetic block from putrescine to spermidine under conditions of massive putrescine over-accumulation. In order to exploit whether the tissue concentrations of spermidine and spermine could be increased, we generated several transgenic mouse lines overexpressing rat Ado- MetDC gene. Among the five lines produced, none displayed an increase in AdoMetDC activity nearly as dramatic as seen in ODC overexpressing animals. The increase in the enzyme activity was at the best fivefold in comparison with their syngenic littermates and the tissue pools of spermidine and spermine of transgenic animals showed only marginal changes [77]. Also, in hybrid mice overexpressing both ODC and AdoMetDC, the tissue levels of the higher polyamines did not differ from those in syngenic mice. Pulse labeling experiments with primary fetal fibroblasts obtained from doubly transgenic mice indicated that polyamine flow was faster in the transgenic than in nontransgenic fibroblasts. AdoMetDC overexpressing ani- mal did not show any phenotypic alterations [77]. One transgenic mouse line likewise generated moderately (two–sixfold) overexpressing human spermidine synthase gene but showing no perturbations in tissue polyamine homeostasis or phenotypic changes [78]. With combined overexpression of ODC and spermidine synthase in hybrid transgenic mice neither brought about any changes in normal polyamine homeostasis [78]. Activation of polyamine catabolism Overexpression of spermidine/spermine N 1 -acetyltransferase Polyamine homeostasis. As indicated earlier (Fig. 1), sper- midine and spermine are converted back to putrescine through the concerted action of SSAT and PAO. PAO is constitutively expressed and strongly prefers acetylated polyamines as the substrates [3] while SSAT is highly inducible, has a very short half-life and serves as the rate- controlling enzyme in polyamine backconversion [4]. The first founder animal overexpressing SSAT (UKU169F 0 ) was a female mouse harboring more than 50 SSAT gene copies in its genome. The animal was extremely small, hairless and infertile. Tissue polyamine pools were dramat- ically distorted. The transgenic brain contained an extremely high concentration of N 1 -acetylspermidine, a compound not normally found in mouse tissues and greatly reduced spermidine pool, while in liver, the putrescine pool was strikingly expanded and that of spermine greatly reduced [79]. As the animal was infertile, no transgenic line could be established. The second founder animal was male that gave rise to two different kinds of offspring, animals harboring only a few SSAT gene copies and retaining their hair (line UKU165a) and animals having more than 20 SSAT gene copies and permanently losing their hair at 3–4 weeks of age (line UKU165b). Members of the UKU165a showed only marginal alterations in their tissue polyamine pools while members of the UKU165b displayed polyamine pool changes typical for SSAT overexpression: large increase in tissue putrescine pool, appearance of N 1 -acetylspermidine and decreases in spermidine and/or spermine pools [79]. Interestingly, these changes developed in the presence of only moderately elevated tissue SSAT activity [79]. Our results (unpublished) have indicated that overexpression of SSAT under these conditions does not decrease the hepatic pool of acetyl-CoA. In an attempt to correct SSAT-induced perturbations in polyamine homeostasis, we generated a hybrid transgenic mouse line overexpressing both ODC and SSAT under the control of mouse metallothionein I promoter. Unexpect- edly, these animals showed much more striking signs of activated hepatic polyamine catabolism than the SSAT overexpressing animals [80]. Even under the condition of severe depletion of spermidine and spermine pools, tremen- dously high tissue putrescine was not driven further to replenish the reduced pool of spermidine. We understand from these results that catabolism is the overriding control mechanism in polyamine metabolism [80]. Regulation of transgene-derived SSAT by polyamine analogues. SSAT is known to be powerfully induced by the higher polyamines and especially by certain polyamine analogues [4]. The regulation of SSAT expression by polyamines and their analogues apparently occurs at many levels of gene expression. These include enhanced transcrip- tion and stabilization of the transcript [81,82], enhanced mRNA translation [83,84] and stabilization of the enzyme protein [85]. The transgenic animals and cells derived from them typically accumulate large amounts of SSAT-specific mRNA that is, however, extremely poorly translated in the absence of polyamines or their analogues [79,86]. This is exemplified by the observation showing that in the pre- sence of N 1 ,N 11 -diethylnorspermine (DENSPM), a power- ful inducer of SSAT, nontransgenic cells display SSAT activity 10· higher than that in transgenic cells not exposed to the analogue but containing 10· more SSAT-specific mRNA [86]. In fact, working with transgenic mice overex- pressing SSAT under the control of mouse metallothionein I promoter, we found striking evidence for a post-tran- scriptional regulation of the transgene expression by DENSPM [87]. In spite of the heterologous promoter, hepatic transgene-derived SSAT was stimulated more than 40 000-fold by the analogue with marginal changes of transcript levels [87]. We proposed that the polyamine analogue could directly interact with SSAT mRNA and improve the translability of the message. It is not excluded that polyamine analogues could alter the splicing of SSAT pre-mRNA as certain viruses appear to induce alternative splicing of the SSAT transcript [88]. Polyamine analogues are potential cancer chemothera- peutic agents and, in fact, DENSPM has undergone clinical 882 J. Ja ¨ nne et al. (Eur. J. Biochem. 271) Ó FEBS 2004 trials [89]. The antiproliferative action of the polyamine analogues is believed to be attributable to an induction of SSAT activity and subsequent depletion of the higher polyamines. This conclusion is based on comparisons between the different inductions of SSAT activity by the analogues and their antiproliferative activity [90–93]. The interpretation of such comparisons between genetically dissimilar cell lines may, however, be difficult as the analogues may have multiple sites of action in different cell lines. We approached the issue by isolating fetal fibroblasts from nontransgenic and SSAT overexpressing mice and exposing the cells to the analogue. We now had a pair of similar cells differing only in the number of SSAT gene copies. We found that the SSAT overexpressing cells were much more sensitive to DENSPM-induced growth inhibi- tion than the nontransgenic cells [86]. Transgenic mice overexpressing SSAT were also more sensitive to the general toxicity of the polyamine analogue [94]. A recent report indicates that small interfering RNA targeted to SSAT mRNA not only prevented SSAT induction by DENSPM, but also prevented apoptosis [95]. As will be shown below, overexpression of SSAT in transgenic animals not only profoundly altered tissue polyamine homeostasis, but likewise created a very complex phenotype affecting skin, fat depots, female fertility, pan- creas, liver and the central nervous system. Skin. As indicated earlier, the first founder animal permanently lost its hair at an early age, as did the members of the UKU165b line harboring more than 20 copies of the SSAT gene [79]. Paradoxically, the hairless skin histology was practically identical to that found in transgenic mice with hair follicle-targeted overexpression of ODC [45], i.e. replacement of hair follicles by large dermal cysts (apparently filled with keratin) and epidermal utriculi, extensive wrinkling of the skin upon aging and lack of subcutaneous fat depots [79,96]. Figure 2 shows a young SSAT transgenic mouse with its syngenic littermate (A) and an old transgenic animal displaying excessive wrinkling of the skin that gives a ÔrhinomouseÕ appearance to the animal (B). The lower panels in Fig. 2 depict the histology of normal skin (C), skin of young (D) and an old (E) transgenic mouse. Note that the normal hair follicles (C) are replaced by dermal cysts (D), which become extremely large in old animals (E). In case of ODC overexpression, the hair loss was attributable to an excessive accumulation of putrescine in the skin, as the loss of hair could be prevented by an early administration of the ODC inhibitor, DFMO [47]. Although indirectly proved, over-accumulation of putres- cine is in all likelihood responsible for the hair loss also observed in SSAT overexpressing mice. This view is supported by experimental findings indicating that putre- scine was constitutively over-accumulated in the skin of these animals and, while the animals properly completed their first hair-cycle, they failed to commence the second anagen phase due to lack of functional hair follicles. Moreover, doubly transgenic mice overexpressing both SSAT and ODC with extremely high levels of putre- scine in the skin displayed distinctly more severe skin histology (significantly larger size of the dermal cysts) than did the singly transgenic mice [96]. Transgenic mice overexpressing SSAT under the control of mouse metallothionein I promoter also lost their hair but much later than those overexpressing the gene under its own promoter [87]. When subjected to the two-stage skin tumorigenesis protocol, SSAT overexpressing mice developed significantly Fig. 2. Hairless phenotype of the SSAT overexpressing mouse. Young SSAT overexpressing mouse with its syngenic littermate (A). An old SSAT transgenic mouse (B). Histology of normal skin showing intact hair follicles (C). Hair follicles are replaced by dermal cysts in SSAT transgenic mouse (D). In old transgenic mouse, the cysts become larger (E) wrinkling the skin. Ó FEBS 2004 Genetic approaches to polyamine functions (Eur. J. Biochem. 271) 883 fewer papillomas than their syngenic littermates [96]. This may be related to the fact that in the syngenic animals, both ODC activity and skin spermidine level increased much more in response to the tumor promoter than in transgenic mice [96]. Coleman et al. [97] employed another approach to study the role SSAT in skin pathophysiology producing, in fact, opposite results. They generated transgenic mice expressing SSAT cDNA under bovine keratin 6 promoter, which directs the expression to the keratinocytes of the hair follicle. The animals were phenotypically indistinguishable from their normal littermates and showed normal hair- cycle. The latter may be attributable to the fact that the low SSAT activity of skin extracts was not increased in the transgenic animals [97]. These animals, however, were much more sensitive to two-stage skin tumorigenesis, as judged by tumor incidence and multiplicity, than their syngenic littermates and showed distinctly enhanced SSAT activity and increased putrescine and N 1 -acetylspermidine level in the papillomas. Interestingly, cysts, derived from dilated hair follicles, were found in the vicinity of the papillomas but less abundantly elsewhere [97]. The obvious inconsistency between present [97] and previous observations [96] may be relatedtothedifferentlevelsofSSAT expression, genetic background or hairlessness. However, mice used in our studies [96] were of a BalbC · DBA/2 background. Mice of DBA/2 are reportedly more sensitive to tumor promotion than animals from a C57BL/6 background [97]. Moreover, all the existing experimental data seem to indicate that an activation of polyamine catabolism is more closely related toantiproliferativeactionthantogrowthpromotion. Female reproductive organs. Histopathological examina- tion of the SSAT overexpressing mice revealed that of 18 tissues and organs examined only skin and female repro- ductive tract were affected in the transgenic animals. Female members of the transgenic line UKU165b were infertile, their uteri were hypoplastic due to a thinner muscular layer and stromal and glandular development was greatly reduced. Examination of the ovaries revealed the presence of primary and small secondary follicles, but absence of larger developing follicles and corpus luteum [79]. Differ- ential display analysis of a gene expression profile of uterus and ovary indicated that the expression of lipoprotein lipase and glyceraldehyde-3-phosphate dehydrogenase was eleva- ted in transgenic animals [98]. Both enzymes are involved in energy metabolism and may have detrimental effects on myometrium and cell viability when overexpressed [98]. SSAT overexpression was also associated with induced expression of insulin-like growth factor binding protein-2 in the uterus and ovary and decreased expression of insulin- like growth factor binding protein-3 in the uterus. These changes may also contribute to the uterine hypoplasia and ovarian hypofunction [98]. It is interesting to note, that ODC overexpression leads to male infertility while SSAT overexpression results in female infertility. Pancreas. Pancreas is the richest source of spermidine in the mammalian body and displays the highest molar ratio of spermidine to spermine, nearly 10 [99,100]. The exact function of such a high spermidine concentration in the pancreas is not known, but may be related to the intense protein synthesis that occurs in this organ. Pancreatic growth appears to be dependent on polyamine biosynthesis, as DFMO retards the growth of the pancreas [101], but does not inhibit the secretory function of the exocrine part of the organ [102]. The cellular functions of the polyamines in the pancreas were approached by generating transgenic rat lines overexpressing SSAT under the control of heavy metal- inducible metallothionein I promoter [103]. Although transgenic pancreas displayed all the signs of activated polyamine catabolism, such as massive accumulation of putrescine and appearance N 1 -acetylspermidine, the levels of the higher polyamines were relatively well maintained. Zinc induction of the promoter resulted in a striking stimulation of the pancreatic SSAT activity in the transgenic animals, but not in the syngenic animals, that was accompanied by an almost total depletion of pancreatic spermidine and spermine and development of histologically verified acute necrotizing pancreatitis [103]. The possibility that pancreatitis would have been caused by reactive oxygen species generated by the action of PAO was excluded by experiments in which PAO was specifically inhibited before zinc administration, showing that the latter inhibition did not alleviate, but rather worsened the pancreatitis [103]. A further piece of evidence causally relating the development of pancreatitis to the profoundly depleted pancreatic polyamine pools came from experiments showing that the inflammatory process could be totally prevented, as judged by histopathology and plasma a-amylase activity, by a prior administration of a-methylspermidine, a metabolically stable spermidine derivative [104]. The results indicated that the higher polyamines are required for the maintenance of metabolic and structural integrity of the pancreas. Induction of SSAT as a cause of acute pancreatic inflammation may have wider applications, especially concerning drug-induced pancreatitis. Gossypol, a cotton seed-derived male antifer- tility agent [105], is known to induce SSAT expression in canine prostate cells [106]. We recently showed that the drug activates polyamine catabolism in the pancreas of normal rats and induces acute pancreatitis through a profound depletion of polyamine in transgenic rats overexpressing SSAT [107]. It thus appears meaningful to screen drugs known to induce pancreatitis for their effect on pancreatic polyamine catabolism. Liver. Polyamines are intimately associated with the growth of mammalian cells. One of the first animal models demonstrating this involved regenerating rat liver after partial hepatectomy. Partial hepatectomy is known to cause an early induction of ODC in the regenerating liver remnant [108,109] followed by a sequential accumulation of putres- cine and spermidine with a slight decrease in spermine [110]. Even though attempts have been made to pharmacologi- cally inhibit rat liver regeneration through blocking ODC, the results have been conflicting [111,112]. Partial hepatec- tomy of transgenic rats expressing metallothionein promo- ter-driven SSAT dramatically induced the enzyme at 24 h after the operation that consequently depleted the hepatic spermidine pool by 80%. As judged by a number of proliferation indicators, the transgenic rats failed to initiate liver regeneration in striking contrast to their syngenic littermates [113]. Only when hepatic spermidine concentra- tion was increased to the preoperative level (apparently due to very high ODC activity) at day 3 after the operation, liver 884 J. Ja ¨ nne et al. (Eur. J. Biochem. 271) Ó FEBS 2004 regeneration slowly commenced in the transgenic animals [113]. The view that the delayed initiation of liver regener- ation in the transgenic animals was causally related to the depletion of hepatic spermidine and spermine pools was strongly supported by experiments revealing that a-methyl- spermidine given prior to the partial hepatectomy fully restored the regeneration [104]. As indicated earlier, SSAT overexpressing mice were extremely sensitive to the toxic effects exerted by polyamine analogues. Treatment of transgenic mice overexpressing SSAT under the control of metallothionein promoter with DENSPM effectively depleted hepatic polyamine pools and resulted in marked mortality that was associated with ultrastructural changes in liver, most notably mitochondrial swelling [87]. Central nervous system. In comparison with the ODC transgenic mice, overexpression of SSAT resulted in even greater expansion of putrescine pool in different regions of the brain. In situ hybridization analyses of the transgenic mice indicated that SSAT was overexpressed in all brain tissue [114]. Some experimental work appears to link enhanced SSAT activity to neuronal damage. This view is based on the finding indicating that kainate-induced seizure activity resulted in an early stimulation of SSAT activity in rat brain [115]. Neurotoxicity of kainate is mediated by a Ca 2+ -dependent process and the drug is particularly toxic to pyramidal cells of hippocampal and cortical neurons [116]. Overexpression of SSAT appears to protect the transgenic animals from kainate-induced toxicity. This was manifested as a substantially reduced (50%) overall mortality of the transgenic mice, in comparison with their syngenic litter- mates, in response to high-dose kainate [114]. The trans- genicity likewise offered a distinct neuroprotection exhibited as a reduced expression of glial fibrillary acidic protein, an commonly used marker of neuronal injury, and no loss of hippocampal neurons in response to kainate in the transgenic animals in comparison with wild-type mice [114]. These results support our earlier view suggesting that expanded pools of brain putrescine, irrespective whether derived from ODC or SSAT overexpression, have a distinct neuroprotective role. The SSAT overexpressing mice likewise showed a significantly elevated threshold, in comparison with their syngenic littermates, to pentylenetetrazol-induced seizure activity involving both tonic and clonic convulsions [117]. Although pentylenetetrazol principally induces epilepsy-like seizure activity through the inhibition of c-aminobutyric acid (GABA), the major inhibitory neurotransmitter [118], a number of reports indicate that antagonists of the NMDA receptor elevate the seizure threshold to pentylentetrazol [119,120]. Interestingly, the difference of seizure threshold to pentylenetetrazol between the transgenic and wild-type animals totally disappeared when ifenprodil, a known NMDA receptor antagonist acting at the polyamine site of the receptor [121,122], was administered prior to pentylenetetrazol [117]. In addition to the elevated seizure threshold, SSAT overexpression likewise protected the animals from pentylenetetrazol-induced neuron loss in the hippocampus [117]. These results are clearly in line with the notion that grossly elevated putrescine levels or the greatly increased (up to 40-fold) molar ratio of putrescine to the higher polyamines in the transgenic brain creates a partial NMDA receptor blockade [117]. Transgenic expression of ODC antizyme As mentioned earlier, antizyme not only inhibits ODC activity, but also facilitates proteasomal degradation of ODC protein and represses polyamine transport. In fact, antizyme occurs in at least three isoforms (antizyme1–3) [16,123]. Unlike antizyme 1, which both inhibits ODC activity and facilitates the degradation of the enzyme protein, antizyme 2 appears to lack the latter function [16]. As mentioned earlier, antizyme 3 is only expressed in testis during late spermatogenesis [35]. Antizyme was shown to be translocated into nucleus during embryonic develop- ment [20] and the protein contains two independent nuclear export signals [21]. Moreover, antizyme forms a ternary complex with the transcription factor Smad1 and protea- somal subunit HsN3 that is translocated into nucleus in response to bone morphogenetic protein receptor activation [22]. In the nucleus, this complex further recruits CBP/p300 repressor SNIP1 and is degraded [23]. A recent review [123] also suggests, based on so far unpublished report, that cyclin D1 and its associated kinase cdk4 interact with antizyme and are degraded in proteasome in a antizyme-dependent fashion. These findings may indicate that ODC antizyme is a general targeting protein for proteasomal degradation. Some recent observations likewise suggest that antizyme possesses functions unrelated to the polyamine metabolism. Antizyme expression is up-regulated in melanoma cells in response to interleukin-1 [124] and antizyme levels are reduced in certain experimental cancers [125]. Antizyme seems to play a specific role in mammalian prostate. Spermine has been identified as an endogenous growth inhibitor in human prostate [126] and it inhibits the growth of poorly metastatic, but not of highly metastatic, rat prostate carcinoma cells [127]. The failure of spermine to inhibit the latter cells is believed to be attributable to the inability of the polyamine to induce antizyme in the highly metastatic cells [127]. If antizyme has functions beyond the metabolism of the polyamines, especially if it interacts with the key players of the cell cycle control, caution should be exercised in interpreting experimental results showing antizyme-dependent growth inhibition. Targeted antizyme 1 expression has been achieved in several transgenic mouse models. The structural part of transgene construct used has been a mutated rat antizyme cDNA where a single nucleotide deletion eliminates the need for frameshifting in translation [128]. Cardiac hypertrophy Two transgenic mouse lines were generated constitutively overexpressing mutated antizyme cDNA under the control of cardiac a-myosin heavy chain promoter targeting the expression to the heart [128]. Even though antizyme effectively inhibited cardiac ODC activity, some residual activity was left and the changes in polyamine pools were small with no changes in cardiac function [128]. A prolonged exposure of syngenic mice to isoprenaline elevated cardiac ODC activity, significantly expanded putrescine and spermidine pools and increased cardiac Ó FEBS 2004 Genetic approaches to polyamine functions (Eur. J. Biochem. 271) 885 growth. Identical treatment of transgenic mice did not activate cardiac polyamine biosynthesis and their tissue accumulation, but induced similar cardiac hypertrophy as seen in wild-type mice [128]. The result was somewhat unexpected as earlier studies have indicated that a specific inhibition of ODC by DFMO can prevent b-adrenergic agonist-induced cardiac hypertrophy [72,129]. Skin tumorigenesis Antizyme expressionhasalsobeentargetedintoskinwith bovine keratin 5 and keratin 6 promoters in several lines of transgenic mice in order to study the role of ODC in the two-stage skin tumorigenesis. In comparison with syngenic mice, the transgenic mice displayed greatly reduced epidermal and dermal ODC activity and spermidine content in response to tumor promotion [130]. All the transgenic lines showed decreased susceptibility to develop papillomas in response to the two-stage chemical carcin- ogenesis protocol [130]. Although earlier studies have convincingly shown that a specific inhibition of ODC by DFMO inhibits skin tumorigenesis in this model [131,132], the present approach is more specific as the inhibition of ODC activity by the antizyme occurs in skin cells and not all over the body. Gastrointestinal carcinogenesis Chemical carcinogenesis in the fore-stomach of zinc-defici- ent mice is another model where the role of ODC has been studied using targeted expression of the antizyme. Antizyme expression significantly reduced both tumor incidence and multiplicity in response to N-nitrosomethylbenzylamine, promoted apoptosis and reduced the expression of cyclin D1 and cdk4 in the fore-stomach of the transgenic mice [133]. The view that the reduced tumor incidence was related to ODC inhibition and not to direct effects of antizyme on some cell cycle regulators [123], was strongly supported by experiments indicating that inhibition of ODC by DFMO reduced tumor incidence and promoted apoptosis in a similar fashion as did transgenic expression of antizyme [133]. Based on these and the skin tumorigenesis studies [130], the authors propose that antizyme may represent a tumor suppressor gene. Gene disruption technology applied to the enzymes of polyamine metabolism Many of the genes of the polyamine metabolizing enzymes have been disrupted either in transgenic mice or mouse embryonic stem cells. In addition to targeted disruption of the genes, there is a X-linked dominant mutation in mice that involves a genomic deletion containing spermine synthase gene. The following paragraphs list the existing knowledge of disruption of genes involved in polyamine metabolism. Arginase II Arginase enzyme, degrading arginine to ornithine and urea, occurs in two isoforms, cytosolic arginase I, which partici- pates in the urea cycle, and mitochondrial arginase II, which apparently is involved in the polyamine synthesis [134]. Mice with a targeted disruption of arginase II gene have been created recently. Homozygous arginase II-deficient mice were viable and otherwise indistinguishable from wild- type mice, except for showing significantly elevated plasma arginine levels. Polyamine analyses of several tissues (brain, liver, kidney and testis) did not reveal any differences between mutant mice and their wild-type counterparts [135]. Although the deficiency in arginase II appears to be a benign trait under normal conditions, it is possible that this deficiency could be deleterious under certain pathophysio- logical conditions. ODC Studies employing inhibitors of polyamine biosynthesis have indicated that an inhibition of ODC will arrest murine embryonic development at the morula-blastocyst stage and an inhibition of AdoMetDC at an even earlier stage [136]. Moreover, DFMO induces resorption of murine embryos when given just after the first week of gestation [137,138]. Studies with mice harboring a disrup- ted ODC gene have revealed that heterozygous animals were viable and fertile while homozygous embryos underwent implantation and induced maternal deciduali- zation, but failed to develop further. This was apparently due to marked apoptosis occurring in the pluripotent cells of the inner cell mass shown as substantial DNA breakage [139]. The fact that ODC-deficient embryos developed to the blastocyst stage, i.e., to more advanced stage than those grown in vitro in the presence of DFMO [136], was in all likelihood attributable to maternal components [139]. Attempts to rescue the embryos through supple- mentation of the pregnant females with putrescine were unsuccessful, apparently due to the toxicity of the diamine (high diamine oxidase activity). As to the reasons for lethality of ODC-deficient embryos, the latter authors offer two possibilities: oxidative DNA damage in the absence of polyamines and inhibition of DNA methyla- tion due to excessive accumulation of decarboxylated AdoMet in the absence of putrescine [139]. Similar ODC gene disruption in the nematode Caenor- habditis elegans results in a virtually normal phenotype when grown in complex medium [140], but when the ODC- deficient nematodes were transferred into polyamine-free medium they showed a phenotype strongly affecting oogenesis and embryogenesis [141]. AdoMetDC As in the case of ODC, homozygous AdoMetDC deficiency is not compatible with murine embryogenesis while hetero- zygous animals were viable, normal and fertile [142]. AdoMetDC-deficient embryos developed normally to the blastocyst stage, but died shortly thereafter or during the early stage of gastrulation at the latest. They developed distinctly further than did embryos cultured in the presence of an inhibitor of AdoMetDC, methylglyoxal bis(guanyl- hydrazone) [136]. When cultured in vitro, AdoMetDC-defi- cient blastocysts showed an absolute growth requirement for spermidine [142]. Unlike ODC-deficient blastocysts [139], AdoMetDC deficiency did not result in DNA fragmentation at the blastocyst stage [142]. The mouse 886 J. Ja ¨ nne et al. (Eur. J. Biochem. 271) Ó FEBS 2004 [...]... replacement of the natural polyamines from their intracellular binding sites In any event, these studies created the important observation indicating that SSAT activity is absolutely necessary for the conversion of spermidine to putrescine, but not for the conversion of spermine to spermidine, paving the way to the discovery of a specific spermine oxidase We have recently also generated SSAT-deficient mice, the. .. addition to these drugs affecting polyamine biosynthesis and/or functions, spermine deficiency also sensitized the cells to the antiproliferative actions of etoposide, an inhibitor of topoisomerase II inducing single and double strand breaks in DNA [147] The drug inhibited the growth of spermine-deficient cells time- and dose-dependently much more effectively than that of the parental cells [145] The generation... early lethality of the decarboxylase-deficient embryos is, however, far from clear They may be related directly to the polyamines – oxidative stress or inadequate synthesis of nucleic acids and proteins in the absence of the polyamines – or, in the case of ODC deficiency, an excessive accumulation of decarboxylated AdoMet – inhibiting DNA methylation and thus disruption of the programming of the embryonal... than the parental cells to the antiproliferative effect exerted by a polyamine analogue However, the observed resistance was not directly related to the depletion of cellular polyamines, as even in the absence of any SSAT activity the intracellular polyamine pools were as effectively depleted as in wild-type cells [6] This may indicate that the depletion of intracellular polyamine pools by polyamine... studies using pharmacological interventions, perturbations of polyamine homeostasis are remarkably difficult to achieve Overexpression of ODC leads to enhanced accumulation of putrescine, yet the diamine is not converted to the higher polyamines One way to force the accumulation of the higher polyamines could be a hybrid mouse overexpressing ODC and lacking SSAT activity Testes appear to be one of the target... an activation of their catabolism in human colonic mucosa [161] Even though these results are supported by a number of experimental works suggesting that an inhibition of ODC in combination with nonsteroidal anti -in ammatory drugs can inhibit colon carcinogenesis [164,165], it is clearly too premature to assign an exclusive role for the polyamines in the development of intestinal tumors the X-chromosome,... members [166] The symptoms were linked to the ability of spermine to serve as a gateway molecule for inward rectifier K+ channels, a function not fully taken over by spermidine [166] SSAT Another X-linked rare syndrome called keratosis follicularis spinulosa decalvans (KFSD) affecting the skin and eye may also be related to the metabolism of polyamines The syndrome apparently includes a duplication of a region... disruption of the spermine synthase gene has been accomplished in mouse embryonic stem cells In the total absence of spermine, the targeted cells grew at a rate that was practically similar to that of the parental cells and displayed no morphological abnormalities [145] The latter may be related to the fact that spermine deficiency resulted in a compensatory increase in cellular spermidine content, in all... characterization of which is underway Table 1 summarizes the phenotypic changes resulting from the genetic modifications of the enzymes involved in polyamine metabolism Clinical consequences of mutations in genes involved in polyamine metabolism There are a few examples of clinical conditions attributable to mutations, polymorphism or rearrangements of genes participating polyamine metabolism ODC Skin-targeted... not the G-allele On the other hand, others have found that the ODC promoter derived from the minor A allele more effectively drives the reporter gene than that from the major G allele [163] Aspirin was without effect on ODC but activated SSAT The apparent synergism between the polymorphism of the ODC gene and the use of aspirin was proposed to be a combined result of an inhibition of polyamine biosynthesis . with induced expression of insulin-like growth factor binding protein-2 in the uterus and ovary and decreased expression of insulin- like growth factor binding protein-3 in the uterus. These changes. biosynthesis, as DFMO retards the growth of the pancreas [101], but does not inhibit the secretory function of the exocrine part of the organ [102]. The cellular functions of the polyamines in the pancreas were. regards the physiological functions of the polyamines, however, experimental evidence exists that assigns specific roles to the polyamines in general and to individual polyamines in particular. The polyamines,